If I am trying to classify 230k text abstracts, which option would be better and more precise when aplying LDA?

  • $\begingroup$ You should try all the options and analyze/evaluate the results imho $\endgroup$ – Erwan Oct 18 '19 at 12:41
  • $\begingroup$ Happen to know how to evaluate when different algorithms are being used? Is it ok if I use perplexity on both? $\endgroup$ – Melania CB Oct 18 '19 at 16:21
  • $\begingroup$ technical measures like perplexity are not suitable for a proper evaluation because they are not based on on the semantics of the task. If a proper evaluation is not possible they might give a partial indication though. See e.g. datascience.stackexchange.com/q/60506/64377 $\endgroup$ – Erwan Oct 19 '19 at 12:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.